Multiple enzymes degrade extracellular ATP in brain tissue, but only Nt5e degrades AMP to adenosine. (A Upper) HPLC-UV analysis at 260 nm identified inosine (INO), AMP, ADP, adenosine (ADO), and ATP (each 10 μM). (Lower) HPLC-UV analysis of media samples collected over time (0, 5, 15, 30, and 60 min) from brain slice incubated with ATP. ATP accumulated into ADP, AMP, adenosine, and inosine linearly (r² > 0.99) over time at rates of 2.80 ± 0.76, 2.60 ± 0.13, 0.60 ± 0.26, and 0.24 ± 0.12 nmol/min per slice, respectively (representative experiment). Notably, spontaneous degradation of ATP, ADP, and AMP in media was insignificant in the absence of slices (ATP, 0.057 ± 0.006%; ADP, 0.004 ± 0.003%; AMP, 0.000 ± 0.000% free phosphate accumulation per minute, mean ± SEM). (B) Catabolism of ATP in slices in the presence of ARL 67156 or in slices from CD39^−/− mice, both in the presence of AOPCP. Bars denote percentage of accumulated ADP or AMP from ATP compared with control (mean ± SEM, n = 6–9 slices). (C) Formation of adenosine from AMP from wild-type control, Nt5e^−/− mice, wild type with AOPCP, and wild-type control slices without AMP. Bars denote percentage of accumulated adenosine compared with control (mean ± SEM, n = 9 slices). (D) Formation of free phosphate from AMP in similar experimental setup as in C (mean ± SEM, n = 9 slices). *P < 0.05; ***P < 0.001; Student t test, all compared with wild type.

Endogenous A1 receptor activation during cortical seizures in vivo is not a consequence of cellular ATP release. (A) Focal seizures were induced in adult mice by placing a sodium penicillin crystal on the exposed cortical surface. The local spread of hyperexcitability was monitored by two extracellular recording electrodes placed at a distance of 1.5 and 2.5 mm from the site of seizure initiation. (B) Representative traces from wild type, wild type with DPCPX, A1R^−/−, Nt5e^−/−, and wild type with AOPCP. The position of the dotted line in the middle of the gray box indicates the average latency of 11.8 ± 1.7 min (mean ± SEM) between application of sodium penicillin and arrival of seizure activity at first recording electrode site in wild-type mice (n = 9). (C) Histogram comparing the latency (minutes) in the five groups (*P < 0.05, ANOVA with Newman–Keuls test, n = 9 animals). (D and E) Histograms comparing maximal frequency and maximal amplitude of neuronal spiking activity detected within the 30-min observation period after application of sodium penicillin (frequency, P = 0.24; amplitude, P = 0.10, ANOVA, n = 9 animals).

A1-mediated synaptic depression in the hippocampus is not a consequence of cellular ATP release and precedes astrocytic ATP release. (A) Heterosynaptic depression was induced in hippocampal CA1 neurons by using two electrodes: one stimulating electrode to evoke eEPSPs in one afferent pathway (100 μA for 100 us every 10 s) and the HFS electrode to deliver HFS in a separate location that was independent of the first afferent pathway (200 μA 100 Hz for 1 s, 100 pulses). (Scale bars: vertical, 4 mV; horizontal, 4 ms.) (B) HFS induced Ca²⁺ wave propagation in Rhod-2 loaded astrocytes (red). The patched neuron is loaded with Alexa 488 (green). [Scale bars: 40 dF/F (%); image, 100 μm.] (C) The P2 receptor antagonists, PPADS or suramine, partly inhibited Ca²⁺ wave propagation. (D) HFS induced heterosynaptic depression in the presence of DPCPX, AOPCP, PPADS, and suramin. (E) HFS induced heterosynaptic depression as evidenced by a depression in the eEPSPs (filled circles; n = 7, mean ± SEM). Representative eEPSP traces are shown in A. (F) Histogram shows effect of AOPCP in wild-type slices (n = 6) or slices from Nt5e^−/− mice (n = 6) with or without DPCPX (n = 5) on depressed. Student t test, *P < 0.05; **P < 0.01.

Selectively activating astrocytic calcium does not affect excitatory transmission. (A) Whole-cell recordings in CA1 pyramidal neurons with eEPSPs evoked every 10 s. (B Upper panel) Astrocytes loaded with the Ca²⁺ indicator Rhod-2/am (3 μM) (red) and caged Ca²⁺ (NP-EGTA/am, 10 μM) was targeted by UV flash, which elicited a propagating Ca²⁺ wave that passed the impaled neuron (green). (Scale: 50 μm.) (Lower panel) Traces represent Ca²⁺ changes over time in individual cells as shown in Upper panel. (C) The eEPSP amplitude before photolysis and during the propagation of the Ca²⁺ wave was used for detecting local increases in extracellular adenosine concentration. The astrocytic Ca²⁺ waves failed to suppress the amplitude of the eEPSP (mean ± SEM, n = 7, Student t test).

Selectively activating excitatory neurons triggers release of adenosine and synaptic depression. (A) Selective stimulation of CA1 hippocampal neurons and targeting of adenosine metabolism. The impaled CA1 neuron was stimulated by injecting the minimal current needed to induce repeated action potentials, and synaptic depression was assessed by comparing eEPSP amplitude before and immediately after depolarization. (B) A train of 30 depolarization pulses in the impaled neuron (green) failed to trigger Ca²⁺ increases in surrounding Rhod-2/am loaded astrocytes (red). (C) Average Ca²⁺ increases in astrocytes (n = 11). (D) Representative trace of eEPSPs before and after (indicated by dotted boxes) the 1-s depolarization. (E Upper) Representative eEPSP traces before (black line) and after (gray line) the train of action potentials in the presence of DPCPX, in slices prepared from Nt5e^−/− mice, and when inosine (a competitive substrate for the equilibrative nucleoside transporters) was added to the pipette solution. (Lower) Summary of several eEPSP traces (n = 5–11, ANOVA before repeat firing, P = 0.1). (F Upper) Number of action potentials elicited during stimulation (n = 5–11). (Lower) Percentage change in eEPSP amplitude between before and after stimulation. **P < 0.01, ANOVA. (Scale bar: 50 μm.)